Plume de nom: An Experimental Approach to Scribal Hand Identification using OCR Technology
Antonia Karaisl completed her PhD at the Warburg Institute in 2020; her thesis on Enlightenment philosopher Christian Wolff’s pre-modern welfare state model was published as a monograph with De Gruyter in 2021. More recently, she has been working on the transcription and translation of the astronomical books of the Fons Memorabilium, an encyclopaedia compiled by humanist Domenico Bandini. In 2016, she co-founded Rescribe, a small not-for-profit company developing bespoke Optical Character Recognition (OCR) solutions for historic books and medieval manuscripts. With the aim of making OCR technology more accessible, Rescribe provides open source software and disseminates new research by publishing in journals and attending conferences.
Antonia’s project combines her interest in medieval and Renaissance palaeography with her technological background: during her fellowship at I Tatti, she will conduct a series of experiments on neural network-based OCR models with the aim of testing their analytical capacity for identifying the handwriting of chosen humanists. Typically, OCR is employed for the automatic transcription of digitized documents, rather than analytical purposes. There have been previous attempts to train artificial intelligence systems to identify medieval script types but the task at hand is complex and by-in-large these attempts have not been all too straightforwardly successful. Antonia’s project is experimenting with a new approach to this open question, testing a hypothesis sprung from Rescribe’s previous research on OCR for Carolingian minuscule. Rather than taking the whole range of medieval hands as a baseline, the chosen area of experimentation will be deliberately small: famous Humanists‘ hands from the 14th-16th century. The actual inquiry will involve close consultation of academic sources on the topic of humanist handwriting in order to combine scholarly perspectives with technological possibilities. Although the scope of the project is deliberately restricted, the chief intent is to validate the hypothesis behind a prospective methodology which, if successful, could serve broader purposes.